In cross-couplings of dihaloarenes, the step(s) that take place at the end of one catalytic cycle—after reductive elimination—and before the start of the next (before oxidative addition) influence the selectivity for mono- versus difunctionalization. When palladium is supported by a bulky ancillary ligand “L”, a competition exists between a second oxidative addition (leading to difunctionalization) and bimolecular displacement of Pd0 from the monofunctionalized product by a second smaller ligand. Because oxidative addition of aryl bromides is faster than aryl chlorides, more difunctionalization might be expected with dibromoarenes compared to dichloroarenes. However, the opposite has been reported in some Suzuki-Miyaura cross-couplings. Here, we investigate the mechanistic origin of this counterintuitive trend to explain why dibromoarenes can sometimes be less prone to diarylation, despite undergoing faster oxidative addition than their chlorinated analogues. The selectivity outcome is found to be closely tied to solvents: the discrepancy between the expected and actual selectivity with dibromoarenes primarily appears in oxygen-containing solvents of at least moderate polarity (like THF), whereas high selectivity for diarylation can be achieved in most aromatic and chlorinated solvents. The results suggest that, in polar oxygen-containing solvents, bromide anion (formally dissolved KBr, the byproduct of oxidative addition) displaces LPd0 from the mono-cross-coupled product as anionic [BrPd0L]–. Based on product distributions in these solvents, the rate of this process is competitive with the rate of intramolecular oxidative addition en route to the diarylated product. In contrast, the analogous displacement of Pd0 by chloride (the byproduct when using dichloroarenes) appears to be a much slower process, if it occurs at all.